Cloudification is immensely important to CSPs because of its potential utility, usability, portability, scaleability,and performance across the world's transformed networks. However, network software is not yet available in cloud native form and probably won't be for quite some time yet. Thus a series of technological stepping stones are needed for CSP's to be able to cross to the sun-kissed uplands of the cloud native landscape and NFV is one of those islands in the stream. As such it may well be in place for a while.
Indeed, as Oliver Cantor says, it is inevitable that operators will have to manage complex, hybrid legacy SDN/NFV plus cloud native environments for years to come, as will those running private business and enterprise networks. Indeed, according to Oliver Cantor, business and enterprise networks comprise some of the most complex and "dirty" environments of all and will pose very significant challenges as the industry moves to a software-defined future. That can't be achieved in one single movement, the evolution is a logical progression to eventually to reach cloud native including SDN, NFV, the management and orchestration of such network, the care and maintenance of pre-existing physical networks and the difficult shift to containerisation.
Oliver Cantor says CSPs are putting themselves through all the traumas associated with total network transformation for one overarching reason - agility. He says customers are demanding increased adroitness across their networks and so the classical supply model of the telco sector where cost issues dominate everything that a CSP does is being replaced by another model where, while cost is still a major factor, speed is of the absolute essence. And increased speed is achieved primarily via software.
As Oliver Cantor observes, "A network has to be a software defined agile beast that is not tethered to its platform or to the architecture underneath." To ensure this can happen standards bodies (of which there are probably too many) will have to come together and collaborate and to at least agree on what framework should be constructed to get software into the new network, otherwise, too much time and money will be spent on building proprietary versions, many of which will be too restrictive, reactionary and, in the longer term, simply not relevant.
Filmed at SDN NFV World Congress, The Hague, 2019